U.S. patent application number 14/012124 was filed with the patent office on 2014-03-06 for methods and apparatus for determining cuff blood pressure.
This patent application is currently assigned to Board of Trustees of Michigan State University. The applicant listed for this patent is Board of Trustees of Michigan State University. Invention is credited to Jin-Oh Hahn, Jiankun Liu, Ramakrishna Mukkamala.
Application Number | 20140066793 14/012124 |
Document ID | / |
Family ID | 50188455 |
Filed Date | 2014-03-06 |
United States Patent
Application |
20140066793 |
Kind Code |
A1 |
Mukkamala; Ramakrishna ; et
al. |
March 6, 2014 |
METHODS AND APPARATUS FOR DETERMINING CUFF BLOOD PRESSURE
Abstract
A method is provided for determining blood pressure for a
subject using a sphygmomanometer. The method includes: measuring an
oscillometric cuff pressure waveform of the subject using the
sphygmomanometer; representing the measured waveform with a
physical model accounting for mechanics of the cuff, an artery and
coupling between the cuff and the artery; determining the model
unknowns from the measured waveform; and determining blood pressure
for the subject using the determined model.
Inventors: |
Mukkamala; Ramakrishna;
(Okemos, MI) ; Liu; Jiankun; (Lansing, MI)
; Hahn; Jin-Oh; (College Park, MD) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Board of Trustees of Michigan State University |
East Lansing |
MI |
US |
|
|
Assignee: |
Board of Trustees of Michigan State
University
East Lansing
MI
|
Family ID: |
50188455 |
Appl. No.: |
14/012124 |
Filed: |
August 28, 2013 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61693812 |
Aug 28, 2012 |
|
|
|
Current U.S.
Class: |
600/494 |
Current CPC
Class: |
A61B 5/02225 20130101;
A61B 5/02208 20130101; A61B 5/02 20130101; A61B 5/02241 20130101;
A61B 5/02255 20130101 |
Class at
Publication: |
600/494 |
International
Class: |
A61B 5/022 20060101
A61B005/022 |
Goverment Interests
GOVERNMENT RIGHTS
[0002] This invention was made with government support under Grant
No. 0643477 awarded by the National Science Foundation and U.S.
Army Grant No. W81XWH-10-2-0124. The U.S. Government has certain
rights in this invention.
Claims
1. A method for determining blood pressure for a subject using a
sphygmomanometer having a cuff and a pressure sensor integrated
therein, comprising: measuring an oscillometric cuff pressure
waveform of the subject using the sphygmomanometer, where the
oscillometric cuff pressure waveform indicates the time evolution
of the pressure inside the cuff during inflation and deflation of
the cuff; representing the measured waveform with a physical model,
where the model accounts for a nonlinear relationship between a
cross-sectional area of an artery and blood pressure; determining
unknowns of the physical model using the measured oscillometric
cuff pressure waveform; and determining blood pressure for the
subject using the determined model.
2. The method of claim 1 wherein the model accounts for the cuff,
the artery, including the nonlinear relationship, and coupling
between the cuff and the artery.
3. The method of claim 2 wherein a portion of the model pertaining
to the cuff accounts for a nonlinear relationship between the
oscillometric cuff pressure and volume of air pumped into the
cuff.
4. The method of claim 2 wherein the model of the coupling between
the cuff and the artery ignores compressibility of arm tissue.
5. The method of claim 1 wherein the model has inputs of a blood
pressure waveform and volume of air pumped into and out of the cuff
and an output of the oscillometric cuff pressure waveform.
6. The method of claim 1 further comprises determining the unknowns
of the model using a priori measurements on the cuff and the volume
of air pumped into and out of the cuff.
7. The method of claim 6 further comprises determining the
cross-sectional area of the artery from a priori measurements on
the cuff and the volume of air pumped into and out of the cuff.
8. The method of claim 7 wherein determining the cross-sectional
area of the artery further comprises representing the blood
pressure waveform and the relationship between the cross-sectional
area of an artery and blood pressure in terms of unknown
parameters.
9. The method of claim 8 further comprises estimating the unknown
parameters by optimally predicting the cross-sectional area of the
artery.
10. The method of claim 9 further comprises determining a blood
pressure waveform from the estimated parameters.
11. The method of claim 1 further comprises measuring a blood
volume waveform of the subject using a photoplethysmography; and
determining the blood pressure of the subject based in part on the
measured blood volume waveform.
12. The method of claim 7 wherein determining the cross-sectional
area of the artery further comprises detecting an envelope of a
plot of the cross-sectional area of the artery in relation to
negative oscillometric cuff pressure waveform; and shifting the
envelope so that its peak derivative is at zero trans-mural
pressure, thereby yielding the relationship between the
cross-sectional area of an artery and blood pressure, wherein the
model is non-parametric.
13. The method of claim 12 further comprises determining a blood
pressure waveform from the cross-sectional area of the artery and
the relationship between the cross-sectional area of an artery and
blood pressure.
14. The method of claim 13 further comprises vertically shifting
the blood pressure waveform so that mean of the blood pressure is
equal to oscillometric cuff pressure at maximal cuff pressure
oscillation.
15. The method of claim 7 further comprises representing the
relationship between the cross-sectional area of an artery and
blood pressure with a parametric model; detecting one or more
envelopes of a plot of the cross-sectional area of the artery in
relation to negative oscillometric cuff pressure waveform;
estimating parameters of the parametric model and one or more blood
pressure values by fitting the parametric model to the
envelope.
16. The method of claim 15 further comprises determining a blood
pressure waveform from the cross-sectional area of the artery and
the relationship between the cross-sectional area of an artery and
blood pressure.
17. A method for determining blood pressure for a subject using a
sphygmomanometer having a cuff and a pressure sensor integrated
therein, comprising: measuring an oscillometric cuff pressure
waveform of the subject using the sphygmomanometer, where the
oscillometric cuff pressure waveform indicates the time evolution
of the pressure inside the cuff during inflation and deflation of
the cuff; representing the measured waveform in terms of a physical
model, where the model accounts for a nonlinear relationship
between a cross-sectional area of an artery and blood pressure, and
a nonlinear relationship between the oscillometric cuff pressure
and volume of air pumped into the cuff; determining the
cross-sectional area of the artery using the physical model and the
measured oscillometric cuff pressure waveform; determining
parameters of the relationship between cross-sectional area of an
artery and blood pressure from the determined waveform and the
measured oscillometric cuff pressure waveform; and determining
blood pressure for the subject using the determined waveform and
the determine relationship.
18. The method of claim 17 further comprises representing the
relationship between the cross-sectional area of an artery and
blood pressure with a parametric model; detecting one or more
envelopes of a plot of the cross-sectional area of the artery in
relation to negative oscillometric cuff pressure waveform;
estimating parameters of the parametric model and one or more blood
pressure values by fitting the parametric model to the
envelope.
19. The method of claim 18 further comprises determining a blood
pressure waveform from the cross-sectional area of the artery and
the relationship between the cross-sectional area of an artery and
blood pressure.
20. A method for determining blood pressure for a subject using a
sphygmomanometer having a cuff and a pressure sensor integrated
therein, comprising: measuring oscillometric cuff pressure waveform
of the subject using the sphygmomanometer, where the oscillometric
cuff pressure waveform indicates the time evolution of the pressure
inside the cuff during inflation and deflation of the cuff;
representing the measured waveform in terms of a physical model,
where the model accounts for a nonlinear relationship between a
cross-sectional area of an artery and blood pressure, and a
nonlinear relationship between the oscillometric cuff pressure and
volume of air pumped into the cuff; determining the cross-sectional
area of the artery using the physical model and the measured
oscillometric cuff pressure waveform; determining
non-parametrically the relationship between cross-sectional area of
an artery and blood pressure from the determined waveform and the
measured oscillometric cuff pressure waveform; and determining
blood pressure for the subject using the determined waveform and
the determined relationship.
21. The method of claim 20 wherein determining the cross-sectional
area of the artery further comprises detecting an envelope of a
plot of the cross-sectional area of the artery in relation to
negative oscillometric cuff pressure waveform; and shifting the
envelope so that its peak derivative is at zero trans-mural
pressure, thereby yielding the relationship between the
cross-sectional area of an artery and blood pressure, wherein the
model is non-parametric.
22. The method of claim 21 further comprises determining a blood
pressure waveform from the cross-sectional area of the artery and
the relationship between the cross-sectional area of an artery and
blood pressure.
23. The method of claim 22 further comprises vertically shifting
the blood pressure waveform so that mean of the blood pressure is
equal to oscillometric cuff pressure at maximal cuff pressure
oscillation.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims the benefit of U.S. Provisional
Application No. 61/693,812, filed on Aug. 28, 2012. The entire
disclosure of the above application is incorporated herein by
reference.
FIELD
[0003] The present disclosure relates to methods and apparatus for
determining cuff blood pressure based on a physical model.
BACKGROUND
[0004] Blood pressure is the pressure exerted by the blood on the
vessel wall. It is a principal vital sign. Various methods, both
invasive and non-invasive, have been developed for its
measurement.
[0005] Invasive methods are employed by penetrating the arterial
wall and are typically restricted to critically ill patients.
Non-invasive methods are much more common and generally involve the
use of an inflatable cuff.
[0006] The standard non-invasive method is auscultation. This
method measures systolic and diastolic pressures (SP and DP) by
occluding a brachial artery with an inflatable cuff and then
detecting the Korotkoff sounds during the deflation period with a
stethoscope while observing the pressure inside the cuff with a
sphygmomanometer. However, this method requires a trained operator
to make the measurement.
[0007] The most widely used automated and non-invasive method is
perhaps oscillometry. This method determines SP, DP, and mean
pressure (MP) using an inflatable cuff, which acts as both an
external pressure applicator and a blood volume sensor. More
specifically, as shown in FIG. 1a, the cuff is likewise placed over
usually a brachial artery and inflated to a supra-SP level (e.g.,
180 mmHg) and then slowly deflated to a sub-DP level (e.g., 50
mmHg). So, during the deflation period, the brachial artery
experiences trans-mural pressures ranging from negative to positive
values. Since arterial compliance changes considerably around zero
trans-mural pressure, the amplitude of the blood volume oscillation
(due to the heart beat) varies greatly. This variation accordingly
alters the amplitude of the resulting pressure oscillation that is
sensed inside the cuff as illustrated in FIG. 1b. The blood
pressure values are then estimated from this oscillometric cuff
pressure waveform (i.e., the waveform indicating the time evolution
of the pressure inside the cuff during inflation and deflation of
the cuff as shown in FIG. 1a).
[0008] The original and most popular blood pressure estimation
method is as follows. First, the oscillometric cuff pressure
waveform is high-pass filtered as shown in FIG. 1b. Then, the
envelope of the high-pass filtered waveform is determined as also
shown in FIG. 1b. Next, since the arterial compliance becomes
maximal when unloaded (i.e., at zero trans-mural pressure), MP is
estimated as the cuff pressure at which the envelope is maximal as
shown in FIG. 1. SP and DP are then estimated as the cuff pressures
at which the amplitude of the envelope is some ratio of its maximum
value. The ratios are fixed to empirically selected values (e.g.,
0.61 before the envelope maximum occurs for SP and 0.74 after the
envelope maximum occurs for DP as shown in FIG. 1) rather than
being specific to the patient at the time of measurement. As a
result, this "fixed-ratio" method is heuristic and can be very
inaccurate. The method may be especially error prone with arterial
stiffening (i.e., decrease in arterial compliance around zero
trans-mural pressure) and changes in pulse pressure (PP=SP-DP).
[0009] Numerous methods have been developed to improve upon the
fixed-ratio method. These methods can be categorized into at least
four groups.
[0010] One group of methods employs more than one cuff. However,
these methods are obviously less practical.
[0011] A second group of methods seeks to obtain a more accurate or
more complete high-pass filtered waveform envelope. However, the
fixed-ratio method is then used to determine the BP values.
[0012] A third group of methods apply methods different from the
fixed-ratio method to estimate the blood pressure values. One
method uses the phase spectrum of the oscillometric cuff pressure
waveform. This method, by itself, cannot estimate MP. In addition,
the method likewise resorts to empirical means to estimate SP and
DP from the phase spectrum variations and may therefore yield no
improvement in accuracy. Another method analyzes the shape of each
beat of the oscillometric cuff pressure waveform. In particular,
the duty ratio is calculated as the ratio of the non-flat duration
of the beat to the entire duration of the beat. This ratio
increases as the pressure applied by the cuff decreases, since
trans-mural pressure increases and more oscillatory components can
be observed. SP and DP are then estimated from the duty ratio using
population statistics. Hence, the method is similarly empirical and
may not improve accuracy.
[0013] A fourth group of methods estimates the entire blood
pressure waveform rather than just DP, MP, and SP. First, DP, MP,
and SP are estimated from the oscillometric cuff pressure waveform.
Then, the cuff pressure waveform is measured at a constant cuff
pressure, which is usually sub-DP (e.g., 60 mmHg). Finally, this
waveform is calibrated with the SP, MP, and/or DP to arrive at an
estimated blood pressure waveform. However, the waveform that is
measured is, in fact, a blood volume waveform. Further, the
arterial compliance is nonlinear. Hence, blood volume is, in
general, not linearly related to blood pressure.
[0014] A more accurate method for automated and non-invasive
measurement of the blood pressure waveform is finger-cuff
photoplethysmography (PPG). This less popular method employs a
finger-cuff with a PPG (which measures a blood volume waveform)
embedded in it and the arterial unloading principle. First, the
cuff is likewise inflated and deflated while measuring the PPG to
yield a finger oscillometric blood volume waveform. Then, the blood
volume at which the artery is unloaded is estimated. One possible
way is to find the average blood volume at which the amplitude of
the oscillometric blood volume waveform envelope is maximal during
the deflation period. Finally, the cuff pressure is continuously
varied so as to maintain this blood volume throughout the cardiac
cycle via a fast, servo-control system. In this way, the cuff
pressure equals the pressure inside the artery. Since the unloaded
blood volume can change (e.g., due to vasomotor tone), it must be
estimated periodically. However, the need for the sophisticated
servo-control system makes this method prohibitively expensive.
Further, the continual unloading of the artery restricts blood flow
to the finger. As a result, subjects often cannot tolerate the
method for very long time periods (e.g., at most on the order of
hours).
[0015] In sum, blood pressure estimation from the oscillometric
cuff pressure waveform is empirical and therefore generally error
prone, while blood pressure measurement via the arterial unloading
principle is expensive and inconvenient. Methods and apparatus are
needed to overcome these limitations and thereby improve blood
pressure measurement.
[0016] This section provides background information related to the
present disclosure which is not necessarily prior art.
SUMMARY
[0017] This section provides a general summary of the disclosure,
and is not a comprehensive disclosure of its full scope or all of
its features.
[0018] A method is provided for determining blood pressure for a
subject. The method includes: measuring the oscillometric cuff
pressure waveform from the subject; representing the measured
waveform with a physical model; determining the model unknowns
using the measured waveform; and determining blood pressure for the
subject using the determined model.
[0019] One exemplary embodiment is a parametric method. The method
includes: measuring the oscillometric cuff pressure waveform from
the subject; representing the measured waveform in terms of the
unknown parameters of a physical model accounting for the cuff and
artery and their coupling; estimating the model parameters from the
measured waveform, the known volume of air pumped into and out of
the cuff, and a priori measurements on the cuff; and determining
blood pressure for the subject using the parameter estimates.
[0020] Another exemplary embodiment is a non-parametric method. The
method includes: measuring the oscillometric cuff pressure waveform
from the subject; representing the measured waveform in terms of a
physical model accounting for the cuff and artery and their
coupling; determining the blood volume or vessel area waveform
based on this model from the measured waveform, the known volume of
air pumped into and out of the cuff, and a priori measurements on
the cuff; determining non-parametrically the blood volume or vessel
area-blood pressure relationship from the determined and measured
waveforms; and determining blood pressure for the subject using the
determined relationship and determined waveform.
[0021] Further areas of applicability will become apparent from the
description provided herein. The description and specific examples
in this summary are intended for purposes of illustration only and
are not intended to limit the scope of the present disclosure.
DRAWINGS
[0022] The drawings described herein are for illustrative purposes
only of selected embodiments and not all possible implementations,
and are not intended to limit the scope of the present
disclosure.
[0023] FIGS. 1A and 1B are a graph illustrating the fixed-ratio
method for estimating diastolic, mean, and systolic blood pressures
(DP, MP, SP) from the oscillometric cuff pressure waveform.
[0024] FIG. 2 is a flowchart providing an overview of an example
method for determining blood pressure for a subject;
[0025] FIG. 3 is a diagram of a physical model of oscillometry;
[0026] FIGS. 4A and 4B are graphs illustrating the determination of
the Arterial P-A Relationship via a non-parametric method;
[0027] FIGS. 5A-5D are graphs depicting the performance of the
non-parametric and fixed-ratio methods, respectively, on simulated
oscillometric cuff pressure waveforms;
[0028] FIG. 6 is a graph illustrating an actual (simulated) blood
pressure waveform and the waveform computed via the non-parametric
method;
[0029] FIG. 7 is a diagram depicting the determination of blood
pressure via a hybrid method without requiring detailed cuff
information;
[0030] FIGS. 8A-8D are graphs depicting the performance of the
second hybrid and fixed ratio methods on simulated oscillometric
cuff pressure waveforms;
[0031] FIGS. 9A and 9B are graphs depicting the performance of the
second hybrid method and the fixed-ratio-method on oscillometric
cuff pressure waveforms from patients; and
[0032] FIG. 10 is a diagram of an example system that implements
methods for determining cuff blood pressure.
[0033] Corresponding reference numerals indicate corresponding
parts throughout the several views of the drawings.
DETAILED DESCRIPTION
[0034] FIG. 2 is a diagram depicting an example method for
determining cuff blood pressure. The present disclosure encompasses
the recognition that cuff blood pressure measurement may be
improved by using a physical model of the underlying phenomena.
More specifically, an oscillometric cuff pressure waveform is
measured at 31 from a subject with a standard arm cuff. The
oscillometric cuff pressure waveform is represented at 32 with a
physical model. The model accounts for the mechanics of the cuff
and artery and their coupling, which may or may not incorporate the
mechanics of the compressible arm tissue. The model unknowns are
determined at 33 from the oscillometric cuff pressure waveform.
Finally, DP, MP, SP and the blood pressure waveform are determined
at 34 using the model. In this way, the estimation of blood
pressure is specific to the subject at the time of measurement and
therefore more accurate.
[0035] Alternatively, an oscillometric blood volume waveform is
measured from a subject with a finger-cuff PPG device, and a
physical model is used to represent the blood volume-blood pressure
relationship. Then, the model is determined from the oscillometric
blood volume and cuff pressure waveforms. Finally, the blood
pressure waveform is determined using the model. The blood pressure
waveform may also be subsequently determined from only the blood
volume waveform using the model. In this way, the blood pressure
waveform is continuously obtained with a finger-cuff PPG device
without the need for invoking the expensive and inconvenient
arterial unloading principle.
[0036] In either case, the determined blood pressure waveform may
be mathematically analyzed so as to also yield other important
physiologic variables, such as cardiac output and the central blood
pressure waveform.
[0037] Exemplary embodiments of the present disclosure are
described below. Further areas of applicability will become
apparent from this description. The description is intended for the
purposes of illustration and is not intended to limit the scope of
the disclosure.
[0038] Several physical models of oscillometry have been developed.
For example, the model of Drzewiecki et al. (see Drzewiecki, G., R.
Hood, and H. Apple. "Theory of the oscillometric maximum and the
systolic and diastolic detection ratios". Ann Biomed Eng. 1994
January-February; 22(1):88-96) is illustrated in FIG. 3. This
example model accounts for the pressure-dependent brachial artery
compliance (Arterial P-A Relationship), the compressibility of air
within the cuff as dictated by Boyle's law (Inflation/Deflation),
and the deformation and stretch of the cuff bladder via a nonlinear
pressure-volume relationship (Cuff Bladder). The arm tissue is
assumed to be incompressible in this model. The inputs to the model
are the brachial artery blood pressure waveform P.sub.a(t) and the
volume of air pumped into and out of the cuff V.sub.p(t). The
output is the cuff pressure P.sub.c(t), which also acts as feedback
to both the blood vessel and the cuff. In this model, the cuff
volume Mt) is defined as the difference between the external sheath
volume V.sub.c(t) and the inside volume contacting the arm
(V.sub.i(t)). Other types of physical models are also contemplated
by this disclosure.
[0039] Arterial P-A Relationship: The cross-sectional area of the
brachial artery A(t) is determined via its trans-mural pressure
(i.e., the difference between blood pressure and cuff pressure
(P.sub.TM(t)=P.sub.a(t)-P.sub.c(t))) according to the following
nonlinear relationship:
A ( t ) = d ln [ aP TM ( t ) + b ] 1 + exp [ - cP TM ( t ) ] ( 1 )
##EQU00001##
where a, b, c, and d are subject-specific parameters at the time of
measurement. The brachial artery area A(t) is linked to the cuff
through the volume of the arm V.sub.i(t) as follows:
V.sub.i(t)=A(t)L.sub.c+V.sub.i0 (2)
where L.sub.c is the length of the arm cuff, and V.sub.i0 is the
initial arm volume for a collapsed brachial artery.
[0040] Cuff Bladder: The cuff pressure P.sub.c(t) is determined by
the external cuff volume, which is the sum of the cuff volume and
arm volume (i.e., Mt)=Mt)+V.sub.i(t)), according to the following
nonlinear relationship:
P.sub.c(t)=E.sub.c{[V.sub.e(t)/V.sub.eo].sup.1/n-1}.sup.n (3)
where E.sub.c is the maximum cuff elastance, V.sub.e0 is the zero
stretch volume of the bladder, and n is a constant of
nonlinearity.
[0041] Inflation/Deflation: The cuff volume V.sub.c(t) is
determined by the cuff pressure P.sub.c(t) and the volume of air
pumped into and out of the cuff V.sub.p(t) according to Boyle's law
as follows:
P.sub.A[V.sub.p+V.sub.c0]=[P.sub.A+P.sub.c(t)]V.sub.c(t) (4)
where P.sub.A is atmospheric pressure, and V.sub.c0 is the initial
air volume in the cuff.
[0042] The two model inputs may be defined in terms of the
following equations:
V p ( t ) = { 81 t 0 .ltoreq. t .ltoreq. 3 245 - 45 ( t - 3 ) / 19
t > 3 and ( 5 ) P a ( t ) = P _ a + A 0 sin ( 2 .pi. f HR 60 t +
.phi. 1 ) + A 1 sin ( 4 .pi. f HR 60 t + .phi. 2 ) ( 6 )
##EQU00002##
where P.sub.a is MP, f.sub.HR is heart rate (HR) in Hz, and
A.sub.0, A.sub.1, .phi..sub.1, and .phi..sub.2 are parameters
defining PP and the waveform shape.
[0043] Some potentially useful alternatives include an Arterial P-A
Relationship with fewer parameters or of a different form and a
more accurate model of P.sub.a(t) (e.g., additional sinusoidal
components). For a given V.sub.p(t) and P.sub.a(t) and set of model
parameter values, the cuff pressure P.sub.c(t) may be computed by
simultaneously solving the above equations for each time instant
using a root-finding algorithm. In this way, the model is able to
mimic the standard oscillometric cuff pressure waveform, which was
its original purpose.
[0044] In one embodiment, a parametric physical model approach may
be used to determine blood pressure. The main idea of this approach
is to determine the unknown parameters of a physical model by
fitting the model to the oscillometric cuff pressure waveform and
to then compute the blood pressure values along with the entire
blood pressure waveform using the determined parameters. For
example, using the physical model of Drzewiecki et al., the
following equation arises after combining Eqns. (1)-(4) and (6) and
re-arranging terms:
A ( t ) = 1 L c { V e 0 [ ( P c ( t ) E c ) 1 / n + 1 ] n - P A P A
+ P C ( t ) [ V p ( t ) + V c 0 ] - V i 0 } = d ln [ a { P _ a + A
0 sin ( 2 .pi. f HR 60 t + .phi. 1 ) + A 1 sin ( 4 .pi. f HR 60 t +
.phi. 2 ) - P c ( t ) } + b ] 1 + exp [ - c { P _ a + A 0 sin ( 2
.pi. f HR 60 t + .phi. 1 ) + A 1 sin ( 4 .pi. f HR 60 t + .phi. 2 )
- P c ( t ) } ] ( 7 ) ##EQU00003##
Here, L.sub.c, V.sub.e0, E.sub.c, n, and V.sub.c0 are determined
from a priori experimentation on the employed cuff; V.sub.p(t) is
the known volume of air pumped into and out of the cuff; P.sub.A is
atmospheric pressure; P.sub.c(t) is measured via a sensor inside
the cuff; f.sub.HR is measured from the oscillations in P.sub.c(t);
and P.sub.a is measured as the P.sub.c(t) at which the maximum
amplitude oscillation occurs. Hence, all of these parameters or
waveforms are known. But, the parameters a, b, c, d, A.sub.0,
A.sub.1, .phi..sub.1, and .phi..sub.2 are patient and time specific
and thus unknown. P.sub.a may also be regarded as an unknown
parameter. Further, the parameter a (and possibly other parameters)
may be fixed to some value, as it has little impact on the Arterial
P-A Relationship.
[0045] Alternatively, a PPG (e.g., placed on the finger) may be
used to obtain a better approximation of P.sub.a(t) than Eqn. (6).
For example, P.sub.a(t) may be approximated from the blood volume
waveform measured with a PPG (u(t)) as follows:
P.sub.a(t)=k.sub.1u(t)+k.sub.2 (8)
where k.sub.1 and k.sub.2 are unknown parameters. Substituting this
equation into Eqn. (7) yields the following equation:
A ( t ) = 1 L c { V e 0 [ ( P c ( t ) E c ) 1 / n + 1 ] n - P A P A
+ P C ( t ) [ V p ( t ) + V c 0 ] - V i 0 } = d ln [ a { k 1 u ( t
) + k 2 - P c ( t ) } + b ] 1 + exp [ - c { k 1 u ( t ) + k 2 - P c
( t ) } ] ( 9 ) ##EQU00004##
The number of unknown parameters in this equation may be reduced by
expressing k.sub.2 in terms of k.sub.1 using measured P.sub.a as
follows:
P _ a = k 1 .intg. 0 T u ( .tau. ) .tau. T + k 2 k 2 = P _ a - k 1
.intg. 0 T u ( .tau. ) .tau. T ( 10 ) ##EQU00005##
where T is the heart period. Note that another advantage of using a
PPG is that the number of unknown parameters is reduced.
[0046] The unknown parameters are estimated by matching both sides
of the second equality of Eqn. (7) or (9) to each other during some
or all of the inflation/deflation period using a least squares
search over a physiologic parameter range. Alternatively, a
two-stage estimation approach may be used in which the numerator
unknowns and denominator unknowns are identified in a sequential
manner. Note that in the high trans-mural pressure regime, the
denominator in the right-hand side of Eqn. (7) or (9) can be
approximated as 1. So, first, the unknown parameters in the
numerator are estimated by matching both sides of the equation
during the high trans-mural pressure regime. Then, these estimated
parameters are substituted in the equation, and the unknown
parameters in the denominator are likewise estimated from the low
trans-mural pressure regime. Other methods for estimating the
parameters also fall within the broader aspects of this
disclosure.
[0047] Finally, P.sub.a(t) is determined via Eqn. (6) or (8), and
SP and DP are then given as the maximum and minimum of
P.sub.a(t).
[0048] In another embodiment, a non-parametric physical model
approach may be used to determine blood pressure. The main idea of
this embodiment is to reduce the assumptions underlying the
physical model by employing non-parametric models of the Arterial
P-A Relationship and the blood pressure waveform and to then
determine these non-parametric models from the oscillometric cuff
pressure waveform. The trade-off in using non-parametric models
instead of parametric ones includes reduced robustness to noise.
For example, using the physical model of Drzewiecki et al., the
following equation arises from the first equality in Eqn. (7):
A ( t ) = 1 L c ( V e 0 { ( P c ( t ) E c ) 1 / n + 1 } n - P A P A
+ P C ( t ) [ V p ( t ) + V c 0 ] - V i 0 ) ( 11 ) ##EQU00006##
According to this equation, A(t) can be solved from the known cuff
parameters and known P.sub.c(t) and V.sub.p(t) waveforms. Then, as
shown in FIG. 4A, the upper or lower envelope (or some average of
the two) of the plot relating A(t) to -P.sub.c(t) is identified to
yield the Arterial P-A Relationship to within a horizontal offset
equal to SP or DP (or some average of the two). Next, as indicated
in FIG. 4B, the Arterial P-A Relationship is exactly determined by
horizontally shifting it so that the peak derivative is located at
zero trans-mural pressure. Thereafter, the resulting Arterial P-A
Relationship is applied to compute P.sub.TM(t) from A(t). Finally,
P.sub.c(t) is added to P.sub.TM(t) to yield P.sub.a(t). SP and DP
are then determined as the maximum and minimum of this waveform. In
this way, the Arterial P-A Relationship and P.sub.a(t) are obtained
without assuming any model (e.g., Eqns. (1) and (6)).
[0049] A key assumption of the above embodiment is that the peak
derivative of the Arterial P-A Relationship is located at zero
trans-mural pressure. This assumption may not always be valid. For
example, the assumption breaks down when the c parameter is very
small, which corresponds to severe arterial stiffening in the
neighborhood of zero trans-mural pressure. One potential solution
is to vertically shift the computed P.sub.a(t) so that its mean
value is equivalent to the P.sub.c(t) at which the maximum
amplitude oscillation occurs.
[0050] The non-parametric method described above (with the vertical
shifting of P.sub.a(t)) was tested using measurements simulated
from the physical model of Drzewiecki et al. More specifically,
P.sub.c(t) was simulated via Eqns. (1)-(6) using the nominal model
parameter values given in Drzewiecki et al. The non-parametric
method and the fixed-ratio method were applied to the simulated
P.sub.c(t) to compute SP and DP, and the SP and DP errors were
determined using the known values of SP and DP. This process was
repeated for a range of values of PP and the c parameter. FIG. 5
shows that the non-parametric method was significantly more
accurate than the fixed-ratio method, especially for low values of
c (i.e., increased arterial stiffening in the zero-transmural
pressure regime) and PP. FIG. 6 shows an example of the
correspondence between the computed (red) and actual (blue) blood
pressure waveforms.
[0051] As mentioned above, one drawback of the non-parametric
method is that the peak derivative of the Arterial P-A Relationship
does not always correspond to zero trans-mural pressure. While
vertically shifting the resulting blood pressure waveform helps to
mitigate this drawback, as shown in FIG. 5, it is an imperfect
solution. Thus, a hybrid physical model approach forms the basis
for yet another embodiment. The main idea of this embodiment is to
combine components of the parametric and non-parametric approaches
so as to eliminate the error introduced by the vertical shifting
while preserving benefits of a non-parametric model and to then
determine the hybrid model from the oscillometric cuff pressure
waveform. For example, using the physical model of Drzewiecki et
al., A(t) is solved from the known cuff parameters and known
P.sub.c(t) and V.sub.p(t) according to Eqn. (11). Then, the
Arterial P-A Relationship is obtained to within a horizontal offset
as described above and shown in FIG. 4A. Next, this relationship is
represented with the parametric model described above as
follows:
A ( t ) = d ln [ a ( SP - P c ( t ) ) + b ] 1 + exp [ - c ( SP - P
c ( t ) ) ] for upper envelope ( 12 ) A ( t ) = d ln [ a ( DP - P c
( t ) ) + b ] 1 + exp [ - c ( DP - P c ( t ) ) ] for lower envelope
( 13 ) ##EQU00007##
Here, A(t) and P.sub.c(t) are known, while a, b, c, d, and SP
and/or DP are unknown. These unknown parameters are determined by
matching both sides of Eqn. (12) and/or (13) to each other during
some or all of the inflation/deflation period using a least squares
search over a physiologic parameter range or any other method known
in the art. Finally, the resulting Arterial P-A relationship is
applied to compute P.sub.a(t) from A(t) as described above.
[0052] The hybrid method described above was tested using
measurements simulated from the physical model of Drzewiecki et al.
The testing procedure was identical to that described above. This
method was able to determine SP and DP and the blood pressure
waveform without error.
[0053] In a fourth embodiment, another hybrid physical model
approach may be used to determine blood pressure. The main idea of
this embodiment is similar to the original hybrid approach but is
designed for the case in which the cuff parameters and volume of
air pumped into and out of the cuff are unknown. For example, using
the physical model of Drzewiecki et al., the original hybrid
approach estimates blood pressure from A(t). However, calculation
of A(t) requires detailed cuff information as indicated in Eqn.
(11). To eliminate the need for this information, two observations
are made. First, the difference in the upper and lower envelopes of
the plot relating A(t) to -P.sub.c(t) is equal to the difference in
the envelopes of the plot relating oscillations in A(t) to
-P.sub.c(t) as shown in FIG. 7. Second, the pressure-volume
relationship of the cuff is nearly linear over a wide range as also
shown in FIG. 7. As a result, the measured and high-pass filtered
P.sub.c(t) and the unmeasured oscillations in A(t) are assumed to
be linearly related (e.g., y(t)=kx(t)+g). Note that this assumption
becomes less tenable with decreasing blood pressure.
[0054] So, first, the difference in the upper and lower envelopes
is detected from the plot relating the high-pass filtered
P.sub.c(t) to -P.sub.c(t). Then, this envelope difference
(.delta.P.sub.c) is represented as the difference in the parametric
models for the Arterial P-A Relationships at systole and diastole
as follows:
.delta. P c = e ln [ a ( SP - P c ( t ) ) + b ] 1 + exp [ - c ( SP
- P c ( t ) ) ] - e ln [ a ( DP - P c ( t ) ) + b ] 1 + exp [ - c (
DP - P c ( t ) ) ] ( 14 ) ##EQU00008##
Here, .delta.P.sub.c and P.sub.c(t) are known, while a, b, c, e=kd,
SP, and DP are unknown. These unknown parameters are determined by
matching both sides of Eqn. (14) to each other during some or all
of the inflation/deflation period using a least squares search over
a physiologic parameter range or any other method known in the art.
Finally, the resulting Arterial P-A Relationship to within a k
scale factor is applied to the high-pass filtered P.sub.c(t) to
compute P.sub.a(t).
[0055] The second hybrid method described above was tested using
measurements simulated from the physical model of Drzewiecki et al.
The testing procedure was identical to that described above. FIG. 8
shows that this method was significantly more accurate than the
fixed-ratio method even for a low DP range. The second hybrid
method was also tested using measurements from 10 cardiology
patients. For each patient, the oscillometric cuff pressure
waveform via a standard arm cuff and reference blood pressure via
an invasive brachial artery catheter were simultaneously measured.
FIG. 9 shows that this method was more accurate than the
fixed-ratio method (which was actually a commercial Omron system
with a proprietary method for estimating DP and SP).
[0056] In a fifth embodiment, a physical model approach may be used
to determine blood pressure from a finger-cuff PPG device. The idea
of this embodiment is to measure the oscillometric blood volume
waveform (which is proportional to A(t)) using the device rather
than computing A(t) as done above. Then, the blood pressure
waveform may be determined with either the parametric approach
(e.g., via Eqn. (7) wherein the term following the first equality
is ignored), non-parametric approach (e.g., as shown in FIG. 4), or
the original hybrid approach (e.g., via Eqns. (12) and (13)) with
the oscillometric blood volume waveform obtained with the PPG
substituted for A(t). Note that the Arterial P-V (blood volume)
relationship is also determined with any of these methods. Hence,
the blood pressure waveform may be subsequently measured using this
relationship without any cuff inflations and deflations. More
specifically, the cuff pressure is held constant, and the waveform
obtained from the PPG is mapped through the relationship to
determine P.sub.TM(t). Then, the constant cuff pressure is added to
this waveform to arrive at P.sub.a(t). A lower constant cuff
pressure will be more tolerable to the subject, while a constant
cuff pressure near MP may be more accurate. Hence, a constant cuff
pressure that is sub-DP may represent a good compromise. Since the
Arterial P-V relationship can change (e.g., with vasomotor tone),
it should be periodically computed (e.g., every 5 to 15 min) via
cuff inflation and deflation to yield the oscillometric blood
volume waveform. In addition, if the subsequent MP is very
different from the MP determined from cuff inflation and deflation,
then the Arterial P-V relationship could be re-computed. In this
way, the blood pressure waveform may be continuously obtained from
a finger-cuff PPG device without invoking the arterial unloading
principle.
[0057] The techniques described herein may be implemented by one or
more computer programs executed by one or more processors. The
computer programs include processor-executable instructions that
are stored on a non-transitory tangible computer readable medium.
The computer programs may also include stored data. Non-limiting
examples of the non-transitory tangible computer readable medium
are nonvolatile memory, magnetic storage, and optical storage.
[0058] Some portions of the above description present the
techniques described herein in terms of algorithms and symbolic
representations of operations on information. These algorithmic
descriptions and representations are the means used by those
skilled in the data processing arts to most effectively convey the
substance of their work to others skilled in the art. These
operations, while described functionally or logically, are
understood to be implemented by computer programs. Furthermore, it
has also proven convenient at times to refer to these arrangements
of operations as modules or by functional names, without loss of
generality.
[0059] Unless specifically stated otherwise as apparent from the
above discussion, it is appreciated that throughout the
description, discussions utilizing terms such as "processing" or
"computing" or "calculating" or "determining" or "displaying" or
the like, refer to the action and processes of a computer system,
or similar electronic computing device, that manipulates and
transforms data represented as physical (electronic) quantities
within the computer system memories or registers or other such
information storage, transmission or display devices.
[0060] Certain aspects of the described techniques include process
steps and instructions described herein in the form of an
algorithm. It should be noted that the described process steps and
instructions could be embodied in software, firmware or hardware,
and when embodied in software, could be downloaded to reside on and
be operated from different platforms used by real time network
operating systems.
[0061] FIG. 10 depicts an example system 110 that implements one or
more of the methods described above. The system 50 is comprised
generally of a sphygmomanometer 111, a model estimation module 114,
a monitor module 116 and at least one output device, such as
display 116. The system may also include a photoplethysmograph 112.
It can be appreciated that other types of sensors may also be part
of the system.
[0062] The sphygmomanometer 111 is configured to measure blood
pressure of a subject. The sphygmomanometer 111 includes a cuff and
a pressure sensor integrated therein as is readily understood in
the art. It is envisioned that the method described above may be
implemented using similar types of pressure meters.
[0063] In operation, the model estimation module 114 is configured
to receive an oscillometric cuff pressure waveform from the
sphygmomanometer 111. The model estimation module 114 is also
configured to receive a measure of the volume of air pumped into
and out of the cuff. From these and other inputs, the model
estimation model 114 can determine parameters of a physical model
that accounts for mechanics of the cuff, an artery and coupling
between the cuff and the artery. The parameters of the model can be
determined by implementing the various methods described above. In
an exemplary embodiment, the model is pre-configured in a data
store of the system and thus accessible to the parameter
determination module 54.
[0064] Given the determined parameter values of the model, the
monitor module 116 can compute the blood pressure waveform of the
subject. In this way, the estimation of blood pressure is specific
to the subject at the time of measurement and therefore more
accurate. In some embodiments, the model estimation module 114 may
also receive a blood volume waveform, for example using a
finger-cuff photoplethysmography. The monitor module 116 can in
turn determine the blood pressure waveform based in part on the
measured blood volume waveform.
[0065] In some embodiments, the monitor module 116 may monitor
computed blood pressure and trigger alarms when monitored
quantities exceed thresholds. In other embodiments, the monitor
module 116 may administer therapy to the subject or modify the
subject's therapy, based on the monitored blood pressure. Lastly,
the monitor module 116 may interface with the display 118 to
present the monitored blood pressure on the display 118. However,
it can be appreciated that other types of output devices may be
used in lieu of the display device.
[0066] This system 110 may be specially constructed for the
required purposes, or it may comprise a general-purpose computer
selectively activated or reconfigured by a computer program stored
on a computer readable medium that can be accessed by the computer.
Such a computer program may be stored in a tangible computer
readable storage medium, such as, but not limited to, any type of
disk including floppy disks, optical disks, CD-ROMs,
magnetic-optical disks, read-only memories (ROMs), random access
memories (RAMs), EPROMs, EEPROMs, magnetic or optical cards,
application specific integrated circuits (ASICs), or any type of
media suitable for storing electronic instructions, and each
coupled to a computer system bus. Furthermore, the computers
referred to in the specification may include a single processor or
may be architectures employing multiple processor designs for
increased computing capability. In other embodiments, the term
module can refer to an application specific integrated circuit
(ASIC), an electronic circuit, a combinational logic circuit,
and/or other suitable components that provide the described
functionality.
[0067] The foregoing description of the embodiments has been
provided for purposes of illustration and description. It is not
intended to be exhaustive or to limit the disclosure. Individual
elements or features of a particular embodiment are generally not
limited to that particular embodiment, but, where applicable, are
interchangeable and can be used in a selected embodiment, even if
not specifically shown or described. The same may also be varied in
many ways. Such variations are not to be regarded as a departure
from the disclosure, and all such modifications are intended to be
included within the scope of the disclosure.
* * * * *